Looking at Students and P2P -- With Data

ORLANDO -- Between the deluge of litigation from the entertainment industry and defiant opposition from college students, is there a way out of the deadlock surrounding the debate on peer-to-peer file sharing?

Researchers now conducting an in-depth study of students' downloading habits -- and how they respond to policy changes -- think there is, and the magic bullet isn't a call for a restructuring of the industry or a punitive approach to music theft. It's data.

So much data, in fact, that banks of computers at Carnegie Mellon University are still processing the reams of numbers collected last year. Although the full extent of the study's conclusions isn't known yet, those involved in the project shared some results of the data analyzed so far: students' Internet activity during the full month of April, 2007, on Illinois State University's campus network.

The university, a major producer of K-12 teachers and home of the Digital Citizen Project, which funded the study with support from both higher education associations and industry, allowed its IT specialists to monitor the Internet use of all students who live on campus for three separate months: April of last year, September of last year and April of this year. The last two months are still being analyzed.

The unprecedented scope of the project -- complicated by the sheer amounts of data, not to mention the privacy concerns that had to be addressed before multiple institutional review boards -- has allowed, and will continue to allow, researchers an in-depth peek at students' downloading habits. Some of the results -- such as that iPods dominate in students' choices for portable music players or that peak downloading times are around 1 a.m. -- confirm conventional wisdom. Others upend it, while data soon to be released could provide new insights into what kinds of interventions can reduce illegal file sharing on campus.

Perhaps most surprising for some will be the finding that not all students, in fact, are born technological geniuses. Many of them, in fact, aren't aware of the difference between peer-to-peer file sharing (using clients such as LimeWire and Kazaa) and sharing files over instant messaging clients or Facebook. That highlights what the researchers believe is the most important way to approach the problem of illegal downloading: education.

"The one thing we do know is that we cannot assume the students know more than they know," said Warren Arbogast, the founder and president of Boulder Management Group, a consulting firm, who is also on the Digital Citizen Project's management team.

Arbogast and others from the university presented their findings thus far at the annual Educause conference, still under way here in Orlando.

One point Arbogast stresses is that file sharing shouldn't be thought of as a technological issue -- one that can be solved by installing the right monitors or blocking the right addresses. As he pointed out, the end result of that mindset is a kind of "technological whack-a-mole," an arms race in which students and industry (with higher education caught in the middle) are constantly trying to outdo each other with new, innovative ways to share music and video.

Sure, campuses can block peer-to-peer file sharing completely, but then again, students could just as easily swap music videos freely available on YouTube or stream music from perfectly legal Web sites, and even install a browser plug-in to convert that streaming media into formats downloadable to their iPods. In other words, there's always going to be the next method of sharing music, movies and TV, a fact that's been true ever since the advent of cassette mix tapes and VHS.

"I think the bottom line for me is that this ... now is less an IT issue than it is an education issue. We’ve got a problem with theft, and I think the question is do we want to try and do something about it?" said Arbogast in an interview.

To illustrate that point, the study found that in April of last year, 51 percent of students living on campus were detected using the peer-to-peer protocol, legal or not. Forty-two percent of students living on campus transferred material that was copyrighted -- in other words, they were illegally downloading or uploading files. On average, there were six copyrighted titles transferred per student per week.

So that means a lot of students share copyrighted music, movies and TV shows -- but also that a lot don't. The remaining question, Arbogast said, was to investigate the "moment of truth" -- in the seventh or eighth grade, when downloading habits are first formed -- when some students decide to start sharing files and others don't.

To collect the massive amounts of data for the study, the university used what is called deep packet inspection technology through a system called Audible Magic, which can identify the signatures of files being transferred on a network and compare them to known copyrighted works. During the study, the researchers concluded that DPI technology was effective in limited ways -- for example, it cannot detect encrypted peer-to-peer traffic, which students can easily employ with the click of a button.

The technology, they concluded, was a useful diagnostic tool to detect activity and target specific students for educational outreach programs. But if used as part of a more punitive approach, the researchers warned, students could just alter their behavior for a short time or switch to encrypted methods or torrents, which can't easily be tracked. "If you threaten us, we’ll respond," Arbogast said, likening the response to a driver passing a police car. "Not the way you think."

The study began after a period in which the university received hundreds of "takedown" notices issued by the entertainment industry under the Digital Millennium Copyright Act, in addition to four subpoenas. At that moment, Arbogast quipped during the session, he realized "we had a research project on our hands."

Even after data collection ended, the university is learning more about how to respond to students' downloading habits. It now blocks all peer-to-peer file traffic on campus, except for students who read and agree to an online pledge that they will only use it for legal purposes (such as the popular game World of Warcraft, not the smallest source of P2P traffic on many campuses).

The results are striking: After collecting data showing that over half of 5,600 or so on-campus students share files, only 112 students are now signed up for the ability to use the P2P protocol. Of those, two have been caught using the network for illegal purposes. "This really comes down to an education issue and a responsibility issue," Arbogast said.

Othercampuses have tried similar awareness-based interventions, mainly with success. Whether such efforts will catch on may depend on future research. The computers at Carnegie Mellon are still chugging away.